From: Tom Lane <tgl@sss.pgh.pa.us>
To: Erik Jones <ejones@engineyard.com>
cc: pgsql-admin@postgresql.org
Subject: Re: Any potential race condition danger from repeated config entries?
In-reply-to: <8D1F2E9D-2476-4BC4-86D3-A5500970D5ED@engineyard.com>
References: <8D1F2E9D-2476-4BC4-86D3-A5500970D5ED@engineyard.com>
Comments: In-reply-to Erik Jones <ejones@engineyard.com>
 message dated "Tue, 14 Jan 2014 11:39:29 -0800"
Date: Tue, 14 Jan 2014 18:16:53 -0500
Message-ID: <1786.1389741413@sss.pgh.pa.us>
Precedence: bulk
Sender: pgsql-admin-owner@postgresql.org

Erik Jones <ejones@engineyard.com> writes:
> For example, checkpoint_segments set to 10 in the main postgresql.conf and then set to 100 in an file included at the end of the main config results in log messages like the following when a reload is issued:
> 2014-01-14 11:10:22.597 PST,,,22347,,52b8a115.574b,6,,2013-12-23 12:46:13 PST,,0,LOG,00000,"parameter ""checkpoint_segments"" changed to ""10""",,,,,,,,,""
> 2014-01-14 11:10:22.597 PST,,,22347,,52b8a115.574b,7,,2013-12-23 12:46:13 PST,,0,LOG,00000,"parameter ""checkpoint_segments"" changed to ""100""",,,,,,,,,""

> At that point the server is running with a value of 100 for checkpoint segments.  Another reload results in the same two messages.  Is there any danger from it being dropped to 10 and immediately set back to 100?

No.  Any one process is going to read the whole config file(s) before it
does anything with the new values, so the only actual effect of the
earlier setting is the excess bleating in the postmaster log.

			regards, tom lane


-- 
Sent via pgsql-admin mailing list (pgsql-admin@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-admin